What licensing operational data means
Operational data is the record of how work gets done inside your company: the work orders, tickets, approvals, schedules and exceptions that pile up in your software every day. Licensing it means giving an AI data company permission to use an anonymized version of those records for a defined purpose, usually training or evaluating AI models, under an agreement you negotiate.
You are not selling your company's data outright. You keep ownership. The agreement sets out what is shared, how it is anonymized, how it may be used, and how you are paid.
Who is buying, and why now
AI models have already been trained on most of the public internet. What they lack is detail about how real businesses operate: the steps, judgment calls and exceptions that never appear in public text. A cluster of AI data companies now licenses this kind of data from established businesses and supplies it to AI developers.
“Companies are sitting on decades of records that show how real work gets done, and that data is now some of the most valuable material for training and evaluating AI.”
Mercor spokesperson, quoted in Barron's
Our buyer comparison lists the companies currently accepting applications, with their published requirements.
What buyers look for
Every buyer is different, but the published requirements and descriptions share a few themes:
- An established company. Most buyers set a minimum size, commonly 20 to 50 full-time employees, and some require several years in operation.
- Records kept in software. Accounting systems, CRMs, ticketing tools, project trackers and dispatch software produce structured histories that buyers can work with.
- Multi-step work involving judgment. Triage, scheduling around constraints, approvals and exception handling are generally more useful than simple data entry.
- Data you have the right to license. If your records belong to clients under contract, or include regulated information, licensing may not be possible without their consent.
How deals are structured
Terms vary by buyer and by company. A few patterns appear across published programs:
- Workflow and corpus data. micro1, for example, distinguishes between workflow partnerships, which capture how work moves through your tools, and corpus partnerships, which license existing anonymized records.
- Estimates before commitment. Buyers such as Nyne offer an estimate or scoping call before you commit to anything.
- You decide the scope. Handshake AI, for example, says the company determines exactly what is shared.
We don't quote figures. Value depends on how much data you have, how detailed it is, and how hard it would be to find elsewhere. Only the buyer can estimate that after its review.
Privacy and anonymization
Reputable programs remove personal information before data is used: customer and employee names, contact details, addresses, payment details and free-text notes that could identify someone. The details matter, so get the buyer's process in writing and have your counsel review it.
Be especially careful with health records, client files held under professional confidentiality, and anything covered by a customer contract that limits how you use their data.
Questions to ask before you sign
- Exactly which systems, records and date ranges are in scope?
- How is personal information removed, and who verifies it?
- Who sees the raw data before anonymization, and where is it stored?
- Which AI developers will be able to use the data, and for what purposes?
- Is re-identification of people or customers prohibited in the agreement?
- Is the license exclusive or non-exclusive?
- How and when are you paid, and what triggers each payment?
- How do you end the agreement, and what happens to the data when you do?
- What liability does each side carry if something goes wrong?
- Does anything in your customer or vendor contracts restrict this use?
How to choose a buyer
Start with the buyers whose published requirements you meet, then weigh how well your records match what each one describes wanting. If you keep financial or scheduling records in common tools, some buyers name those systems specifically. Apply to your best match first; if it passes, move to the next.
The checker does this for you. Answer four questions and it ranks the buyers whose requirements you meet.
Red flags
- Asking for raw data before an agreement is signed.
- Vague answers about how personal information is removed.
- Upfront fees charged to your company.
- Exclusivity demands without a clear reason or payment for it.
- Pressure to sign before your counsel has reviewed the agreement.
More questions
Do we share data to apply?
No. Applying requires basic company information. Data is scoped and shared only after a buyer reviews fit and you sign an agreement.
Can we license data to more than one buyer?
It depends on the agreement. Some licenses are non-exclusive and some ask for exclusivity. Check the exclusivity terms before you sign.
This guide is general information, not legal advice. Have your counsel review any agreement before you sign.